A software tool for stitching two PET/CT body segments into a single whole-body image set

作者:Chang Tingting; Chang Guoping; Clark John W Jr; Rohren Eric M; Mawlawi Osama R*
来源:Journal of Applied Clinical Medical Physics, 2012, 13(3): 179-189.
DOI:10.1120/jacmp.v13i3.3599

摘要

A whole-body PET/CT scan extending from the vertex of the head to the toes of the patient is not feasible on a number of commercially available PET/CT scanners due to a limitation in the extent of bed travel on these systems. In such cases, the PET scan has to be divided into two parts: one covering the upper body segment, while the other covering the lower body segment. The aim of this paper is to describe and evaluate, using phantom and patient studies, a software tool that was developed to stitch two body segments and output a single whole-body image set, thereby facilitating the interpretation of whole-body PET scans. A mathematical model was first developed to stitch images from two body segments using three landmarks. The model calculates the relative positions of the landmarks on the two segments and then generates a rigid transformation that aligns these landmarks on the two segments. A software tool was written to implement this model while correcting for radioactive decay between the two body segments, and output a single DICOM whole-body image set with all the necessary tags. One phantom, and six patient studies were conducted to evaluate the performance of the software. In these studies, six radio-opaque markers (BBs) were used as landmarks (three on each leg). All studies were acquired in two body segments with BBs placed in the overlap region of the two segments. The PET/CT images of each segment were then stitched using the software tool to create a single DICOM whole-body PET/CT image. Evaluation of the stitching tool was based on visual inspection, consistency of radiotracer uptake in the two segments, and ability to display the resultant DICOM image set on two independent workstations. The software tool successfully stitched the two segments of the phantom image, and generated a single whole-body DICOM PET/CT image set that had the correct alignment and activity concentration throughout the image. The stitched images were viewed by two independent workstations from two different manufacturers, attesting the ability of the software tool to produce a DICOM compliant image set. The study demonstrated that this software tool allows the stitching of two segments of a whole-body PET/CT scan with minimal user interaction, thereby facilitating the interpretation of whole body PET/CT scans from a number of scanners with limited extent of bed travel.

  • 出版日期2012

全文